Molecular Phylogenetics and Evolution 36 (2005) 722–727

www.elsevier.com/locate/ympev

Short communication

Molecular phylogeny of Chromodoris (Mollusca, Nudibranchia)

and the identiWcation of a planar spawning clade
Nerida G. Wilson a,¤, Michael S.Y. Lee b
a
Environmental Biology, The University of Adelaide, SA 5005, Australia
b
The South Australian Museum, North Terrace, Adelaide, SA 5000, Australia

Received 16 February 2004; revised 9 July 2004

Available online 8 June 2005

1. Introduction lesson, 2000) of the optimal trees. Miller (1980) empha-

Chromodoris, Alder and Hancock, 1855, is the most
speciose genus in the nudibranch family Chromodorid-
idae, and has a cosmopolitan distribution with the
greatest diversity occurring at low latitudes. It is esti-
mated to contain approximately 200 species with up to
22% of known species still not oYcially described (Gos-
liner and Draheim, 1996); presumably many remain to
be discovered. Some species have shown unique chemi-
cal defenses (Gavagnin and Fontana, 2000) and future
screening would beneWt from an evolutionary perspec-
tive. A phylogeny will also facilitate analysing colour
evolution, as within Chromodoris there appear
instances of both convergent (due to mimicry: Rud-
man, 1991) and homologous colour traits (Gosliner and
Behrens, 1998). Currently, there are no phylogenetic
hypotheses for this large and important genus. Many
colour groups were described in Chromodoris to facili-
tate identiWcation using external colour patterns (e.g.,
Rudman, 1983, 1985, 1987), with no expectation that
these groupings reXected phylogenetic relationships.
However, recent work has suggested one of these
groups shares features in internal anatomy and repro-
duction, and may represent a real clade (Gosliner and
Behrens, 1998; Wilson, 2002).
The monophyly of Chromodoris has also not yet been
rigorously tested and cannot be assumed: in phyloge-
netic studies that contained 2–3 Chromodoris species, the
genus appeared monophyletic in only some (Thollesson,
1999; Wollscheid-Lengeling et al., 2001) or none (Thol-
*
Corresponding author. Fax: +61 7 3365 4755.
E-mail address: nerida.wilson@adelaide.edu.au (N.G. Wilson).
sised the diYculty in separating the typically bipolar
genus Cadlina Bergh, 1879 from Chromodoris, and sug-
gested synonymy.
We used mitochondrial DNA sequence data (partial
16S rDNA) to create a phylogeny for a group where
convergent evolution of colour patterns could hinder
phylogenetic reconstruction based on the most widely
used diagnostic characters.
2. Materials and methods
2.1. Taxon sampling
Seventeen species of Chromodoris and two outgroups
were collected and sequenced directly for partial 16S
rDNA. Sequences of Wve other species of Chromodoris
and three Cadlina species were taken from GenBank (see
Table 1). These exemplars from Chromodoris were cho-
sen so that many putative species-groups and divergent
forms were represented. Specimens were identiWed by
comparison to published literature and dissected when
necessary. The sequenced specimens have been deposited
in the Australian and South Australian Museum collec-
tions (Table 1).
Small portions of ethanol-preserved foot tissue were
washed in STE buVer and transferred to 10% Chelex
solution. After adding proteinase-K (10 mg/ml), the solu-
tion was incubated at 50–55 °C for a few hours, with
occasional mixing, then raised to 90 °C, and the cooled
extraction buVered with 10% TE. The 16S rDNA frag-
ments were ampliWed using the primers 16Sar-L and
16Sbr-H (Palumbi et al., 1991). PCR ampliWcations were

1055-7903/$ - see front matter  2005 Published by Elsevier Inc.

doi:10.1016/j.ympev.2005.04.003
 N.G. Wilson, M.S.Y. Lee / Molecular Phylogenetics and Evolution 36 (2005) 722–727 723

Table 1
Species used in this study, asterisk indicates sequence acquired from GenBank
Species Locality Accession number Voucher number
Actinocyclus verrucosus Mooloolaba, Queensland, Australia AY458799 SAM D19274
Cadlina laevis Marstrand, Bohuslän, Sweden AJ225182* —
Cadlina luarna Costa Rica AF430348* —
Cadlina luteomarginata North Atlantic, USA AF249231* —
Cadlinella ornatissima Heron Island, Queensland, Australia AY458802 AM C203859
Chromodoris alternata Port Phillip Bay, Victoria, Australia AY458800 SAM D19281
Chromodoris ambiguus Port Phillip Bay, Victoria, Australia AY458801 SAM D19260
Chromodoris aspersa Mooloolaba, Queensland, Australia AY458813 SAM D19282
Chromodoris collingwoodi North Stradbroke Island, Australia AY731181 SAM D19283
Chromodoris daphne Moreton Bay, Queensland, Australia AY458803 SAM D19284
Chromodoris epicurea Triabunna, Tasmania, Australia AY458804 SAM D19285
Chromodoris geometrica Mooloolaba, Queensland, Australia AY458805 SAM D19286
Chromodoris krohni NE Atlantic, Spain AF249239* —
Chromodoris kuiteri1 Mooloolaba, Queensland, Australia AY458806 SAM D19287
Chromodoris kuiteri2 Great Barrier Reef, Australia AF249240* —
Chromodoris kuniei Heron Island, Queensland, Australia AY458807 SAM D19261
Chromodoris leopardus Mooloolaba, Queensland, Australia AY458808 SAM D19288
Chromodoris lochi Mooloolaba, Queensland, Australia AY458810 SAM D19289
Chromodoris luteorosea Cadiz, Andalusia, Spain AJ225183* —
Chromodoris magniWca Whitsundays, Queensland, Australia AY458811 SAM D19290
Chromodoris purpurea Cadiz, Andalusia, Spain AJ225184* —
Chromodoris quadricolor Red Sea, Egypt AF249241* —
Chromodoris roboi Heron Island, Queensland, Australia AY458814 SAM D19291
Chromodoris splendida Mooloolaba, Queensland, Australia AY458815 SAM D19292
Chromodoris striatella Mooloolaba, Queensland, Australia AY458809 SAM D19293
Chromodoris strigata Heron Island, Queensland, Australia AY458816 SAM D19294
Chromodoris tasmaniensis Triabunna, Tasmania, Australia AY458817 SAM D19295

carried out using whole genomic preparations. The ther-
mal cycling conditions were an initial denaturation step
at 94 °C for 3 min, followed by 39 cycles of 30 s at 94 °C,
30 s at 55 °C, 1 min at 72 °C, extension for 5 min at 72 °C,
and soaked at 25 °C for 2 min. The PCR product was gel
puriWed and sequenced directly in both directions using
ABI automated sequencing.
2.2. Alignment and analyses
The sequences were reconciled using Sequencher
(Gene Codes, Ann Arbor, MI) and aligned in ClustalX
(Thompson et al., 1997). Initially, a sensitivity analysis
(Giribet et al., 2001) was performed, where several align-
ment parameter combinations were used (gap opening
penalty 2–15, gap extension cost 1–6.66; and transition
weight 0.5–1), and each resultant alignment analysed
(using parsimony for computing speed). The alignments
and trees were all very similar, suggesting that the inXu-
ence of alignment variation was small. One of these
alignments (using default parameters in Clustal, i.e.,
15:6.66:0.5) was chosen for more detailed analysis and
small manual adjustments were made by eye.
The alignment was imported into MacClade (Maddi-
son and Maddison, 2000) and then PAUP* (SwoVord,
2002) for phylogenetic analyses. Hypervariable regions
(69 sites) that were considered unalignable were
excluded from analysis, leaving 416 sites (178 variable,
124 parsimony-informative). The total alignment was
485 sites long, and has been deposited in TreeBASE
(www.treebase.org/). Sequences have been deposited in
the NCBI GenBank database (see Table 1).
All trees were rooted using Actinocyclus verrucosus
and Cadlinella ornatissima, which are related to, but out-
side, a putative Chromodoris-Cadlina clade (Gosliner
and Johnson, 1994; Rudman, 1984). Parsimony analyses
employed heuristic searches with 500 random sequence
additions, with gaps treated as missing data. Clade sup-
port was calculated using bootstrapping and decay indi-
ces. Parsimony-uninformative characters were excluded
from MP bootstraps (10,000 replicates) following discus-
sion in Felsenstein (2004). Branch support (Bremer,
1988) was calculated using Tree Rot v2b (Sorenson,
1999).
The most appropriate model of DNA substitution for
these data was chosen using Modeltest v3.06 (Posada and
Crandall, 1998). Both hierarchical likelihood-ratio tests
(HLTs) and the Akaike Information Criterion converged
on the TVM+I+G model (Wrst named in Posada, 2004).
Bayesian MCMC analyses using MrBayes v3 (Ronquist
and Huelsenbeck, 2001) used the most similar available
model (erring on the side of simplicity): HKY+I+G
(Hasegawa et al., 1985). Four chains (one heated) were
run simultaneously, burn-in was 50,000 generations (well
after likelihood reached stationarity), and the remaining
1,000,000 trees sampled every 100 trees.
724 N.G. Wilson, M.S.Y. Lee / Molecular Phylogenetics and Evolution 36 (2005) 722–727

3. Results and discussion
Uncorrected pairwise distance (alignment–ambigu-
ous sites excluded) ranged from 0.013 between the two
specimens of Chromodoris kuiteri to 0.228 (between
Chromodoris alternata and Chromodoris roboi). The MP
analyses resulted in a single optimal tree (Fig. 1), with
three major clades discussed below. The Bayesian analy-
sis also returned a very similar tree (Fig. 2), but con-
tained minor topological diVerences that aVected nodes
poorly supported in both analyses.
These results indicate that Chromodoris is paraphyletic.
The most basal ingroup clade is reasonably supported
(branch support 5, bootstrap 74, and Bayesian 76) and
contains Chromodoris ambiguus and C. alternata as well as
included cadlinid species, i.e., Cadlina luarna, Cadlina
laevis, and Cadlina luteomarginata. The second clade
(branch support 8, bootstrap 96, and Bayesian 100) con-
tains the type species Chromodoris magniWca and all the
included members of the Chromodoris quadricolor colour
group, as well as Chromodoris aspersa. This clade is here-
after termed the ‘planar spawner’ clade, in reference to the
Xat egg mass laid by its members (Wilson, 2002). While
the clade itself is strongly supported, 16SrDNA provides
little resolution for its internal relationships. MP analyses
suggest C. aspersa as the most basal taxon while Bayesian
analyses indicate Chromodoris lochi. The third major clade
(branch support 4, bootstrap 68, and Bayesian 100) con-
tains the remaining Chromodoris species from southeast-
ern Australia, the Atlantic/Mediterranean, and the Indo-
PaciWc, and is termed the ‘erect spawner’ clade after their
upright egg masses. Many of these subclades are also
diagnosed by unusual morphological and behavioural fea-
tures, as discussed below, although the polarity of these
traits needs to be rigorously assessed.
3.1. Cadlina and the monophyly of Chromodoris
The monophyly of Chromodoris (sensu Rudman) is not
supported by this study. Constraining Chromodoris to be
monophyletic results in an MP tree only six steps longer,
but one that is signiWcantly worse according to both Tem-
pleton’s nonparametric test and the Kishino–Hasegawa
(K–H) test (p D0.014 for both) (see Felsenstein, 2004). The

Fig. 1. Parsimony phylogram based on 16S rDNA for Chromodoris species. Closed box indicates acquisition of planar egg mass. The numbers to the
right represent Bremer support; numbers to the left of each branch represent bootstrap percentages. Bootstrap values are only shown for clades pres-
ent in the majority-rule consensus tree. Branch lengths reconstructed under ACCTRAN optimisation in PAUP*. CI: 0.52, RI: 0.61, and L D 517.
 N.G. Wilson, M.S.Y. Lee / Molecular Phylogenetics and Evolution 36 (2005) 722–727 725

grouping of the southern temperate taxa C. ambiguus and
C. alternata with Cadlina is inconsistent with the current
generic classiWcation. Despite the fact that the included
Cadlina species are from the northern hemisphere, and the
two Chromodoris species are from the Southern Ocean, the
clade is reasonably supported. C. ambiguus, C. alternata,
and most Cadlina species are restricted to very cold waters,
and this physiological trait appears generally characteristic
of this clade. The grouping of these temperate Chromodoris
species with Cadlina is consistent with Rudman’s (1987)
observation that the exogenous sperm sac in C. ambiguus
and C. alternata is the same as that seen in some Cadlina
species. However, Miller’s (1980) proposal that Cadlina be
synonymised with Chromodoris is not the only taxonomic
solution, since an alternative is to transfer C. ambiguus and
C. alternata to another genus.

Fig. 2. Bayesian MCMC majority-rule consensus tree for Chromodoris species based on 16S rDNA and gap information, and the substitution model
(HKY+I+G) most similar to the model selected using HLTs. Numbers represent the posterior probabilities.
726 N.G. Wilson, M.S.Y. Lee / Molecular Phylogenetics and Evolution 36 (2005) 722–727
3.2. Planar spawners
The planar spawning clade is strongly supported and
species are predominately found in the tropical Indo-
West PaciWc. The best trees that lacked this clade were
eight steps longer and these four trees were signiWcantly
worse than the MP tree according to both Templeton’s
nonparametric test and the K–H test (p < 0.05). The clade
contains all included members of the C. quadricolor col-
our group, described by Rudman (1977, 1982). It also
includes C. aspersa that has been linked with the C. quad-
ricolor group previously based on the shared planar egg
masses (Wilson, 2002). A report of C. magniWca laying an
upright egg mass (Klussman-Kolb and Wägele, 2001) is
erroneous, and further observation has shown that
C. magniWca lays a planar egg mass like others in this
clade (Wilson, pers. obs). Planar spawning is arguably a
derived reproductive character diagnosing this clade,
since Actinocyclus, Cadlina, and the other Chromodoris in
this analysis (where known) lay erect egg masses, although
Cadlinella lays planar egg masses (Boucher, 1983).
3.3. Erect spawners
This clade consists of the majority of species of
Chromodoris, excluding the planar spawners and the close
relatives of Cadlina. There were no previous existing
hypotheses of phylogenetic relationships within this clade.
In both analyses, the erect spawners are divided into four
geographically coherent subclades: southeastern Austra-
lian endemics (Chromodoris splendida, Chromodoris
daphne, Chromodoris tasmaniensis, and Chromodoris
epicurea), Atlantic/Mediterranean taxa (Chromodoris lute-
orosea, Chromodoris krohni, and Chromodoris purpurea),
and two clades containing Indo-PaciWc forms (Chromod-
oris collingwoodi and Chromodoris kuniei) (Chromodoris
leopardus, C. roboi, and Chromodoris geometrica). The lat-
ter two subclades, excluding C. collingwoodi, exhibit the
unusual behavioural trait of rhythmically lifting their
mantles. Relationships between these four clades are
labile: the Australian clade is basal in the MP analyses,
while the Atlantic/Mediterranean clade is basal in the
Bayesian analysis. Indo-PaciWc species occur in both the
planar and upright spawning clades of Chromodoris. This
contrasts directly with the phylogenetic structure of
another chromodorid genus, Hypselodoris, where the two
major clades correspond to the Indo-PaciWc and the
Atlantic/Eastern PaciWc regions (Gosliner and Johnson,
1999).
3.4. Colouration, convergence, and common descent
The diversity and similarity of colour patterns in
chromodorid nudibranchs have been attributed in part
to warning colouration (Edmunds, 1991), mimicry (Rud-
man, 1991), or phylogenetic constraints (Gosliner and
Behrens, 1998; Johnson and Gosliner, 1998). The current
phylogeny provides some suggestions as to how these
processes might be responsible for particular aspects of
colouration. The production of pure black pigment
exhibits strong phylogenetic conservativism: it occurs in
all planar spawners with the exception of C. aspersa, and
is absent from almost all erect spawners (present only in
C. geometrica). Other aspects of colour patterns appear
to correlate with presence of black pigment, with the pla-
nar species all showing longitudinal black stripes or
swathes of black. C. aspersa, the only member of the pla-
nar clade that lacked black pigment, also lacked any lon-
gitudinal striped pattern. C. geometrica, the lone species
from the erect spawning clade with black pigment, had a
diVerent (reticulated) pattern.
Background body colour also shows strong phyloge-
netic conservatism within the erect spawners. The Medi-
terranean clade all share pink bodies, while the
southeastern Australian endemics have white bodies,
regardless of pattern. A white background colour is also
present in the Cadlina clade and may represent the plesi-
omorphic state.
Mullerian mimicry (the evolution of similar warning
colouration in toxic species) has been invoked to explain
nudibranch colour patterns common to geographical
regions (Rudman, 1991). All of Rudman’s examples used
changes in spot size or distribution, mostly in Chromod-
oris in southeastern Australia. The phylogeny here
reveals some evidence of mimicry (orange spots common
to both C. ambiguus and C. tasmaniensis) but also that
most homologous colour traits are restricted to particu-
lar geographical regions. Thus, similarities in both col-
our pattern and geographic distribution within
Chromodoris might be explained by phylogeny, rather
than by convergence. Indeed, when faced with similarly
patterned taxa in the same locality, in the absence of a
rigorous phylogeny the most reasonable null hypothesis
would be that they are all descended from a common
ancestor which exhibited that colour pattern and lived in
that locality, rather than invoking dispersal of unrelated
(and diVerently coloured) lineages into the locality and
subsequent convergent evolution of colouration.
Detailed data on geographic co-variation in colour pat-
terns of putative models and mimics in a phylogenetic
framework are required to provide solid proof of
mimicry (Greene, 1997).
Acknowledgments
We thank Dan and Dave Jackson, Wayne Ellis, and
Shane Litherland for Weld assistance, and Andrew
Hugall and Dan Jackson for technical advice. Permis-
sion for collecting was given by the Great Barrier Reef
Marine Park Authority, Qld Parks and Wildlife Services,
the Department of Natural Resources and Environment
 N.G. Wilson, M.S.Y. Lee / Molecular Phylogenetics and Evolution 36 (2005) 722–727
(Vic.) and the Department of Primary Industries, Water,
and Environment, Hobart. Two anonymous referees are
thanked for their suggestions on improving the manu-
script. This research has been supported by the Austra-
lian Research Council (Postgraduate Research Award to
N.W., Research Fellowships and grants to M.L.), and a
University of Queensland Research Grant (to N.W.).
References
Alder, J., Hancock, A., 1855. A Monograph of the British Nudibranch-
iate Mollusca. The Ray Society, London.
Boucher, L.M., 1983. Extra-capsular yolk bodies in the egg masses of
some tropical Opisthobranchia. J. Mollus. Stud. 49, 232–241.
Bremer, K., 1988. The limits of amino acid sequence data in angio-
sperm phylogenetic reconstruction. Evolution 42, 795–803.
Edmunds, M., 1991. Does warning colouration occur in nudibranchs?.
Malacologia 32, 241–255.
Felsenstein, J., 2004. Inferring Phylogenies. Sinauer Associates, Massa-
chusetts (664).
Gavagnin, M.G., Fontana, A., 2000. Dipertenes from marine opistho-
branch molluscs. Curr. Org. Chem. 4, 1201–1248.
Giribet, G., Edgecombe, G.D., Wheeler, W.C., 2001. Arthropod phy-
logeny based on eight molecular loci and morphology. Nature 413,
157–161.
Gosliner, T.M., Behrens, D.W., 1998. Five new species of Chromodoris
(Mollusca: Nudibranchia: Chromodorididae) from the tropical
Indo-PaciWc Ocean. Proc. Cal. Acad. Sci. 50, 139–165.
Gosliner, T.M., Draheim, R., 1996. Indo-PaciWc opisthobranch gastro-
pod biogeography: how do we know what we don’t know?. Am.
Malacol. Bull. 12, 37–43.
Gosliner, T.M., Johnson, R.F., 1999. Phylogeny of Hypselodoris (Nudi-
branchia: Chromodorididae) with a review of the monophyletic
clade of Indo-PaciWc species, including descriptions of twelve new
species. Zool. J. Linn. Soc. Lond. 125, 1–114.
Gosliner, T.M., Johnson, S., 1994. Review of the genus Hallaxa (Nudi-
branchia: Actinocyclidae) with descriptions of nine new species.
Veliger 37, 155–191.
Greene, H.W., 1997. Snakes: The Evolution of Mystery in Nature. Uni-
versity of California Press, Berkeley.
Hasegawa, M., Kishino, K., Yano, T., 1985. Dating the human-ape
splitting by a molecular clock of mitochondrial DNA. J. Mol. Evol.
22, 160–174.
Johnson, R.F., Gosliner, T.M., 1998. The genus Pectenodoris
(Nudibranchia: Chromodorididae) from the Indo-PaciWc, with
the description of a new species. Proc. Cal. Acad. Sci. 50, 295–
306.
Klussman-Kolb, A., Wägele, H., 2001. On the Wne structure of opistho-
branch egg masses (Mollusca, Gastropoda). Zool. Anz. 240, 101–118.
Maddison, D.R., Maddison, W.P., 2000. McClade. Sinauer Associates,
Sunderland, Massachusetts.
Miller, M.C., 1980. Cadlina willani, a new dorid nudibranch (Gastrop-
oda: Opisthobranchia) from New Zealand. New Zeal. J. Zool. 7,
165–172.
727
Palumbi, S., Martin, A., Romano, S., McMillan, W.O., Stice, L., Gra-
bowski, G., 1991. The Simple Fool’s Guide to PCR, Version 2.0.
Department of Zoology and Kewalo Marine Laboratory, Univer-
sity of Hawaii.
Posada, D., 2004. Modeltest v3.4 documentation. http://dar-
win.uvigo.es/software/modeltest.html.
Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA
substitution. Bioinformatics 14, 817–818.
Ronquist, F., Huelsenbeck, J.P., 2001. MrBayes 3: Bayesian Phyloge-
netic Inference Under Mixed Models. http:/morphbank.ebc.uu.se/
mrbayes.
Rudman, W.B., 1977. Chromodorid opisthobranch Mollusca from
East Africa and the tropical West PaciWc. Zool. J. Linn. Soc. Lond.
61, 351–397.
Rudman, W.B., 1982. The Chromodorididae (Opisthobranchia: Mol-
lusca) of the Indo-west PaciWc: Chromodoris quadricolor, C. lineo-
lata and Hypselodoris nigrolineata colour group. Zool. J. Linn. Soc.
Lond. 76, 183–241.
Rudman, W.B., 1983. The Chromodorididae (Opisthobranchia: Mol-
lusca) of the Indo-West PaciWc: Chromodoris splendida, C. aspersa
and Hypselodoris placida colour groups. Zool. J. Linn. Soc. Lond.
78, 105–173.
Rudman, W.B., 1984. The Chromodorididae (Opisthobranchia: Mol-
lusca) of the Indo-West PaciWc: a review of the genera. Zool. J.
Linn. Soc. Lond. 81, 115–273.
Rudman, W.B., 1985. The Chromodorididae (Opisthobranchia:Mol-
lusca) of the Indo-West PaciWc: Chromodoris aureomarginata, C.
verrieri and C. Wdelis colour groups. Zool. J. Linn. Soc. Lond. 83,
241–299.
Rudman, W.B., 1987. The Chromodorididae (Opisthobranchia:Mol-
lusca) of the Indo-West PaciWc: Chromodoris epicura, C. aureo-
purpurea, C. annulata, C. coi, and Risbecia tryoni colour groups.
Zool. J. Linn. Soc. Lond. 90, 305–407.
Rudman, W.B., 1991. Purpose in pattern: the evolution of colour in
chromodorid nudibranchs. J. Moll. Stud. 57, (T.E. Thompson
Memorial Issue, suppl): 5–21.
Sorenson, M.D., 1999. TreeRot. Boston University, Boston, Massachu-
setts.
SwoVord, D.L., 2002. PAUP* version 4b10. Sinauer Associates, Sun-
derland, Massachusetts.
Thollesson, M., 1999. Phylogenetic analysis of dorid nudibranchs
(Gastropoda: Doridacea) using the mitochondrial 16S rRNA gene.
J. Mollus. Stud. 65, 335–353.
Thollesson, M., 2000. Increasing Wdelity in parsimony analysis of dorid
nudibranchs by diVerential weighting, or a tale of two genes. Mol.
Phylogenet. Evol. 16, 161–172.
Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgens,
D.G., 1997. The ClustalX windows interface: Xexible strategies for
multiple sequence alignment aided by quality analysis tools.
Nucleic Acids Res. 25, 4876–4882.
Wilson, N.G., 2002. Egg masses of chromodorid nudibranchs
(Mollusca: Gastropoda: Opisthobranchia). Malacologia 44,
289–305.
Wollscheid-Lengeling, E., Boore, J., Brown, W., Wägele, H., 2001. The
phylogeny of Nudibranchia (Opisthobranchia, Gastropoda, Mol-
lusca) reconstructed by three molecular markers. Org. Divers Evol.
1, 241–256.